【fittype问题】fittype接受的匿名函数中带有阶乘相，MATLAB提示错误

sinpen

先贴代码：

1. t = 1:1:11;
   
2. h = [0.0572 0.0572 0.0572 0.0714 .1143 .1429 .1571 .1429 .1143 .1000 .0714 ];
   
3. plot(t, h, 'or'), hold on
   
4. 
   
5. s = fitoptions('Method', 'LinearLeastSquares');
   
6. 
   
7. ffun = fittype(@(lambda,n)(((lambda.^n)./factorial(n)).*exp(-lambda)),...
   
8.     'Coefficients', {'lambda'},...
   
9.     'Independent', 'n'...
   
10.     );
    
11. 
    
12. [cfun, gof, output] = fit(t', h', ffun,...
    
13.     'StartPoint', [7]);
    
14. 
    
15. plot(cfun,'b')

复制代码

这段script运行到 fittype 函数这里，会提示说传入的函数句柄为不可计算的MATLAB表达式：

2. Error using fittype (line 356)
   
3. Expression (((lambda.^n)./factorial(n)).*exp(-lambda)) is not a valid MATLAB expression, has non-scalar coefficients,
   
4. or cannot be evaluated:
   
5. Error while trying to evaluate FITTYPE function :
   
6. 
   
7. N must be a matrix of non-negative integers.
   
8. 
   
9. Error in poissonFit (line 7)
   
10. ffun = fittype(@(lambda,n)(((lambda.^n)./factorial(n)).*exp(-lambda)),...
    

Error using fittype (line 356)<br /> Expression (((lambda.^n)./factorial(n)).*exp(-lambda)) is not a valid MATLAB expression, has non-scalar coefficients,<br /> or cannot be evaluated:<br /> Error while trying to evaluate FITTYPE function :<br /> <br /> N must be a matrix of non-negative integers.<br /> <br /> Error in poissonFit (line 7)<br /> ffun = fittype(@(lambda,n)(((lambda.^n)./factorial(n)).*exp(-lambda)),...<br />

但是，我直接将这个函数句柄拿出来用，是可以加算的。

sinpen

我深深地怀疑是不是那个factorial阶乘并不适合使用fittype？